github/gallery-dl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

remove '&' from URL patterns

Browse Source 
'/?&#' -> '/?#' and '?&#' -> '?#'

According to https://www.ietf.org/rfc/rfc3986.txt, URLs are
"organized hierarchically" by using "the slash ("/"), question
mark ("?"), and number sign ("#") characters to delimit components"
This commit is contained in:
Mike Fährmann
2020-10-22 23:12:59 +02:00
parent 1686dc1757
commit 968d3e8465
74 changed files with 158 additions and 158 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
gallery_dl/extractor/xhamster.py
View File

@@ -33,7 +33,7 @@ class XhamsterGalleryExtractor(XhamsterExtractor):
"{gallery[id]} {gallery[title]}")
filename_fmt = "{num:>03}_{id}.{extension}"
archive_fmt = "{id}"
pattern = BASE_PATTERN + r"(/photos/gallery/[^/?&#]+)"
pattern = BASE_PATTERN + r"(/photos/gallery/[^/?#]+)"
test = (
("https://xhamster.com/photos/gallery/11748968", {
"pattern": r"https://thumb-p\d+.xhcdn.com/./[\w/-]+_1000.jpg$",
@@ -152,7 +152,7 @@ class XhamsterGalleryExtractor(XhamsterExtractor):
class XhamsterUserExtractor(XhamsterExtractor):
"""Extractor for all galleries of an xhamster user"""
subcategory = "user"
pattern = BASE_PATTERN + r"/users/([^/?&#]+)(?:/photos)?/?(?:$|[?#])"
pattern = BASE_PATTERN + r"/users/([^/?#]+)(?:/photos)?/?(?:$|[?#])"
test = (
("https://xhamster.com/users/goldenpalomino/photos", {
"pattern": XhamsterGalleryExtractor.pattern,